TNF inhibitors are currently considered both effective and cost-effective in patients with active rheumatoid arthritis (RA), particularly in patients who have not responded fully to methotrexate. There is substantial doubt about the cost-effectiveness of TNF inhibitors as initial treatment for active RA. New data from the National Data Bank for Rheumatic Diseases now question the current consensus in methotrexate failures. The data suggest that in routine clinical practice TNF inhibitors provide only modest incremental benefits over best conventional therapy. If confirmed, these observational studies suggest that the economic argument underpinning the widespread use of TNF inhibitors in established RA is unsustainable.